SYSTEM AND METHOD FOR INSPECTING RAILROAD TIES

• US 20130191070A1
• Filed: 03/12/2013
• Published: 07/25/2013
• Est. Priority Date: 06/30/2004
• Status: Active Grant

• Current Assignee: Loram Technologies Incorporated (Loram Maintenance of Way, Inc.)
• Original Assignee: Georgetown Rail Equipment Company (Loram Maintenance of Way, Inc.)
• Inventors: Belcher, Jeb E., Villar, Christopher M., Kainer, John J., Aaron, Charles W., Grissom, Gregory T., Mauricio, Antonio R., Pagliuco, David M., Wamani, Wilson T., Nagel, John A. II, Orrel, Steven C., Bertilson, Zechariah
• Granted Patent: US 8,958,079 B2
• Time in Patent Office: Days
• US Class Current: 702/167
• CPC Class Codes: B61K 9/00 Railway vehicle profile gauges; Detecting or indicating overheating of components; Apparatus on locomotives or cars to indicate bad track sections; General design of track recording vehicles
• CPC Class Codes: B61K 9/08 Measuring installations for surveying permanent way applications of measuring apparatus or devices for track building purposes E01B35/00; measuring techniques G01
• CPC Class Codes: B61L 23/044 Broken rails
• CPC Class Codes: B61L 23/045 Rail wear
• CPC Class Codes: G01B 11/24 for measuring contours or curvatures
• CPC Class Codes: G01C 3/08 Use of electric radiation detectors
• CPC Class Codes: G01M 99/00 Subject matter not provided for in other groups of this subclass

Abstract

A system for inspecting railroad ties in a railroad track includes a light generator, an optical receiver and a processor. The light generator is oriented to project a beam of light across the railroad track while moving along the railroad track in a travel direction. The optical receiver is oriented to receive at least a portion of the beam of light reflected from the railroad track and configured to generate image data representative of a profile of at least a portion of the railroad track. The processor is configured to analyze the image data by applying one or more algorithms configured to find boundaries of a railroad tie and determine one or more condition metrics associated with the railroad tie.

Claims

1. A system for inspecting railroad ties in a railroad track, comprising:

a. a light generator, oriented to project a beam of light across the railroad track while moving along the railroad track in a travel direction;

b. an optical receiver, oriented to receive at least a portion of the beam of light reflected from the railroad track and configured to generate image data representative of a profile of at least a portion of the railroad track; and

c. a processor, configured to analyze the image data by applying one or more algorithms configured to;

i. find boundaries of a railroad tie; and

ii. determine one or more condition metrics associated with the railroad tie.

2. A system in accordance with claim 1, wherein the processor comprises a machine vision algorithm for finding the boundaries of the railroad tie, the algorithm configured to:

a. identify regions of the image data that are smooth relative to surrounding regions, relative to a smoothness threshold;

b. find at least two edges of a smooth region, oriented approximately perpendicular to the travel direction;

c. identify two of the found edges which are spaced apart by approximately a width of a tie as the first and second edges; and

d. identify the first and second edges as the boundaries of the railroad tie.

3. A system in accordance with claim 2, wherein the machine vision algorithm is further configured to:

a. identify regions that include rails, tie plates and/or fasteners;

b. find third and fourth edges of the smooth region, approximately perpendicular to and adjacent to the first and second edges, located outside of the rail, tie plate and/or fastener region, on opposing sides of the track; and

c. identify the first through fourth edges as the boundaries of the railroad tie.

4. A system in accordance with claim 2, wherein at least a portion of the smooth region inside the tie boundaries is identified as a tie surface area.

5. A system in accordance with claim 2, wherein regions that do not satisfy the smoothness threshold and areas physically above the smooth region are identified as non-tie surfaces.

6. A system in accordance with claim 5, wherein the non-tie surfaces are filtered out, and not considered when determining the condition of the railroad tie.

7. A system in accordance with claim 5, wherein the tie surface area is normalized by a quantity of non-tie surface area within the tie boundaries.

8. A system in accordance with claim 5, wherein ties having a quantity of non-tie surface area above a threshold are identified as ties that cannot be graded.

9. A system in accordance with claim 2, wherein the processor is further configured to:

a. determine a tie-to-tie distance between the tie and a next adjacent tie;

b. compare the tie-to-tie distance to an expected tie-to-tie distance;

c. determine that a tie is missing or covered if the tie-to-tie distance is too large, and/or determine that a phantom tie has been identified if the tie-to-tie distance is too small.

10. A system in accordance with claim 9, wherein the processor is further configured to:

a. identify an initially missed tie by reanalyzing areas in which the tie-to-tie distance is too large, to find edges of smooth areas based upon an adjusted smoothness threshold, and/or determine which adjacent area is a phantom tie by comparing adjacent areas in which the tie-to-tie distance is too small.

11. A system in accordance with claim 2, wherein the processor is further configured to:

a. move ahead to a new portion of the image data, after the tie is identified, to a next expected tie location, based upon an expected tie-to-tie distance; and

b. search the new portion of the image data to identify a new smooth region based upon the smoothness threshold.

12. A system in accordance with claim 11, wherein the processor is further configured to:

a. search forward and backward in the new portion of the image data relative to the next expected tie location, to identify the new smooth region; and

b. increase the smoothness threshold with increased distance from the next expected tie location.

13. A system in accordance with claim 11, wherein the processor is further configured to determine that a tie is missing or covered if no new smooth region is found within about two times the expected tie-to-tie distance.

14. A system in accordance with claim 1, wherein the processor is further configured to apply an algorithm for finding tie plates, the algorithm configured to:

a. find objects that are of tie plate size in a suspected general location where a plate should be;

b. eliminate known non-plate objects based on position and size properties; and

c. apply an edge-finding subroutine to enhance an outside edge of the tie plate.

15. A system in accordance with claim 14, wherein the processor is further configured to separate the tie plate from the tie to accommodate height differential.

16. A system in accordance with claim 14, wherein the edge-finding subroutine is configured to:

a. identify a pixel as a plate pixel in a plate blob;

b. stair-step through pixels to determine if they are connected to the current pixel; and

c. identify the “

connected”

pixels as representing the presence of an edge.

17. A system in accordance with claim 14, wherein the processor is further configured to apply an algorithm for identifying a type of the tie plates, the algorithm configured to:

a. determine a horizontal slope of the tie plate;

b. determine the first and second derivatives of the horizontal slope;

c. identify the tie plate as a spike plate if the slope has a negative first derivative; and

d. identify the tie plate as a rolled/elastic-fastener tie plate if the slope has a negative second derivative.

18. A system in accordance with claim 17, wherein the processor is further configured to:

a. search vertically across the tie plate shape to verify that the tie plate extends across a vertical span of the plate if a rolled/elastic-fastener tie plate was identified;

b. identify a pixel as a plate pixel on an outside of the rolled/elastic-fastener tie plate;

c. stair-step through pixels to determine if they are connected to the identified pixel; and

d. identify the connected pixels as representing the presence of an edge.

19. A system in accordance with claim 14, wherein the processor is further configured to apply an algorithm for measuring tie plate cut, the algorithm configured to:

a. find at least one region inside the outside edge of the tie plate;

b. find at least one region outside the outside edge of the tie plate and atop a tie surface area;

c. determine an apparent tie plate thickness as an elevation difference between the at least one region inside the outside edge of the tie plate and the at least one region outside the outside edge of the tie plate; and

d. subtract the apparent tie plate thickness from a known thickness of the plate to determine the tie plate cut.

20. A system in accordance with claim 1, wherein the processor is further configured to:

a. apply a weight factor to each condition metric of the railroad tie; and

b. compute a condition grade for the tie based on the weighted condition metrics.

21. A system in accordance with claim 20, wherein the condition grade is adjusted by a normalized tie surface area.

22. A system in accordance with claim 20, wherein the tie condition grade, tie location and spacing between ties with a tie condition grade above or below a grade threshold are correlated to create a tie replacement plan.

23. A system in accordance with claim 20, wherein the processor is further configured to:

a. determine GPS and heading information of a position of the railroad tie along a railroad track bed; and

b. adjust the condition grade of the tie based upon a degree of track curvature.

24. A system in accordance with claim 20, wherein the processor is further configured to:

a. locate a rail joint and/or weld near the tie; and

b. adjust the condition grade of the tie for higher degradation rates near the rail joint and/or weld.

25. A system in accordance with claim 1, wherein the condition metrics include at least one of:

tie warpage, tie surface roughness, tie condition, tie entropy, tie heuristics, tie length, tie width, tie skew angle, tie plate cut, differential plate cut, spike height, depth of adzing, broken condition, curvature, tonnage, speed and decay zone.

26. A system in accordance with claim 1, wherein the light generator and the optical receiver are attached to a rail vehicle configured for travel along the railroad track in the travel direction.

27. A system in accordance with claim 1, wherein the processor is further configured to:

a. determine a location of the railroad tie with respect to a railroad track bed; and

b. correlate the tie location with GPS and/or milepost location.

28. A system in accordance with claim 1, wherein the processor is further configured to identify un-ballasted bridge ties, the algorithm configured to:

a. identify regions in the image data in which there are voids or shadows between multiple adjacent ties;

b. determine that a tie-to-tie distance between the multiple adjacent ties is consistent with an expected un-ballasted bridge tie spacing; and

c. identify an un-ballasted bridge location as a contiguous section of ties defined by a first and a last of the multiple adjacent ties.

29. A system in accordance with claim 1, wherein the processor is further configured to identify at least one of a fastener type, the start and end of a road/rail crossing at grade, a switch, switch points, switch ties, switch frogs and joint bars.

30. A system in accordance with claim 29, wherein the processor is further configured to adjust tie counts, and/or compile an inventory of found or missing parts per unit of track length.

31. A system for inspecting railroad ties, comprising:

a. a rail vehicle, configured for moving along a railroad track in a travel direction;

b. a light generator, attached to the rail vehicle, oriented to project a beam of light across the railroad track while moving thereon;

c. an optical receiver, attached to the rail vehicle, oriented to receive at least a portion of the beam of light reflected from the railroad track and configured to generate image data representative of a profile of at least a portion of the railroad track; and

d. a processor, configured to analyze the image data by applying one or more algorithms configured to;

i. find boundaries of railroad ties in the railroad track; and

ii. determine the condition of railroad ties.

32. A system in accordance with claim 31, wherein the processor is further configured to:

a. determine a location of each railroad tie with respect to the railroad track; and

b. correlate the tie location with GPS and/or milepost location.

33. A system in accordance with claim 31, wherein the processor is further configured to apply an algorithm configured to:

a. determine a series of condition metrics of the railroad ties;

b. apply a weight factor to each condition metric; and

c. compute a condition grade for each tie based on the weighted condition metrics.

34. A system in accordance with claim 33, wherein the condition metrics include at least one of:

tie warpage, tie surface roughness, tie condition, tie entropy, tie heuristics, tie length, tie width, tie skew angle, tie plate cut, differential plate cut, spike height, depth of adzing, broken condition, curvature, tonnage, speed and decay zone.

35. A system in accordance with claim 33, wherein the processor is further configured to:

a. determine a location of each railroad tie;

b. determine a spacing between railroad ties with a good or bad condition grade; and

c. correlate the tie condition grade, tie location, and spacing between ties with good or bad tie condition grades to create a tie replacement plan.

36. A system in accordance with claim 33, wherein the processor is further configured to:

a. determine GPS and heading information of a position of the railroad tie along a railroad track bed; and

b. adjust the condition grade of the tie based upon a degree of track curvature.

37. A system in accordance with claim 33, wherein the processor is further configured to:

a. locate a rail joint and/or a weld near the tie; and

b. adjust the condition grade of the tie for higher degradation rates near the rail joint and/or weld.

38. A system in accordance with claim 31, wherein the processor is further configured to identify at least one of un-ballasted bridge ties, a fastener type, the start and end of a rail crossing, a switch, switch points, switch ties, switch frogs and joint bars.

39. A method for inspecting railroad ties, comprising:

a. scanning a set of railroad ties with an optical scanning system to produce image data;

b. analyzing the image data using a processor running a machine vision algorithm;

c. identifying, via the processor, boundaries of railroad ties in the set of railroad ties; and

d. determining, via the processor, condition metrics for each railroad tie in the set of railroad ties.

40. A method in accordance with claim 39, wherein the step of scanning the set of railroad ties comprises scanning the railroad ties with a 3-D optical scanning system, and generating a computer-readable image representative of a profile of the railroad ties.

41. A method in accordance with claim 40, wherein scanning the railroad ties with a 3-D optical scanning system comprises:

a. projecting a beam of light across a railroad track bed that includes the railroad ties;

b. receiving light reflected from the railroad ties with a camera and producing a camera signal; and

c. generating, from the camera signal, the computer-readable image.

42. A method in accordance with claim 39, wherein the condition metrics include at least one of:

tie warpage, tie surface roughness, tie condition, tie entropy, tie heuristics, tie length, tie width, tie skew angle, tie plate cut, differential plate cut, spike height, degree of adzing, broken condition, curvature, tonnage, speed and decay zone.

43. A method in accordance with claim 39, further comprising computing a condition grade for each railroad tie based on a weighted summation of the condition metrics.

44. A method in accordance with claim 39, further comprising:

a. determining a location of each railroad tie in the set of railroad ties with respect to a railroad track bed; and

b. creating a tie replacement plan that is based upon the location of each railroad tie.

45. A method in accordance with claim 39, further comprising:

a. determining a location of each railroad tie in the set of railroad ties with respect to a railroad track bed; and

b. correlating the tie location with GPS and/or milepost location.

46. A method in accordance with claim 39, further comprising:

a. determining GPS and heading information of a position of each railroad tie in the set of railroad ties along a railroad track bed; and

b. adjusting the condition grade of the tie based upon a degree of track curvature.

47. A method in accordance with claim 39, further comprising:

a. locating a rail joint and/or weld near the tie; and

b. adjusting the condition grade of the tie for higher degradation rates near the rail joint and/or weld.

48. A method in accordance with claim 39, further comprising identifying at least one of un-ballasted bridge ties, a fastener type, the start and end of a road/rail crossing at grade, a switch, switch points, switch ties, switch frogs and joint bars

49. A method for grading railroad ties of a railroad track in situ, comprising:

a. scanning a set of railroad ties with an optical scanning system and generating image signals representing the railroad ties;

b. applying, via a computer processor, a machine vision algorithm to the image signals and generating a 3-D profile of the railroad ties that locates tie features;

c. analyzing the 3-D profile via a computer processor to determine a set of condition metrics representing each railroad tie in the set of railroad ties; and

d. computing a condition grade for each railroad tie based on a weighted summation of the condition metrics.

50. A method in accordance with claim 49, further comprising:

a. determining locations of ties with respect to the railroad track;

b. correlating the tie location with GPS and/or milepost location relative to the railroad track; and

c. creating a tie replacement plan that is based upon the GPS and/or milepost location.

51. A method in accordance with claim 49, wherein the condition metrics include at least one of:

tie warpage, tie surface roughness, tie condition, tie entropy, tie heuristics, tie length, tie width, tie skew angle, tie plate cut, differential plate cut, spike height, degree of adzing, broken condition, track curvature, track tonnage, speed, grade crossing, proximity of a rail joint, proximity of a weld and decay zone.

52. A method in accordance with claim 51, further comprising adjusting the condition grade of the tie based upon at least one of the condition metrics.